Jump to content

Talk:WikiScience

Add topic
From Meta, a Wikimedia project coordination wiki

What is the status of this project? Is it moving at all? Could you use some help?

Great idea

[edit]

But I don't know if this is dead....

I propose contributors should be registered and have at least one published peer-reviewed paper in a well known journal.


This has also occured to me

[edit]

I have been thinking about this idea too, and it also seems inevitable to me.

A lot of the structure of such a pedia already exists, as journals publish keywords and full-test searches (as seen on Google scholar) can compare the similarity of two articles without human input.

It does seem important to me that everything be fully referenced, preferably with a link to the pubmed abstract. This reduces the need for peer review as these references come from reviewed journals. Pubmed links would allow users in many institutions (such as universities) to quickly link through to the full text of the source article.

Establishing Article Accuracy and Ruggedness

[edit]

I propose you require a supermajority vote with a minimum of 100 votes to publish or edit an article. Every registered user gets one vote per topic per week.

This idea requires an overhaul to wikimedia's publishing model and you will need to get some open source programmers interested enough to contribute some code-hours to it. Let me flesh this out a little bit so that you get it. Let's say you start an article for "Nonnegative_Integer" or whatever. Once you submit it for publishing, it will be delivered to a "New Article" queue. If the article title hasn't been put onto the "New Article" RSS feed, it goes immediately - otherwise it waits in the queue until it can go. Wikiscience users could subscribe to the feed in their favorite reader, and they would be asked to vote for or against any article(s) they were interested in enough to peruse. Given the nature of how RSS is evolving and projects like Google Reader's Share feature, people could links to their friends for topics they cover, or set up a shared voting rss feed that projects you have contributed to come to. The point is you could interest a lot of people in voting, if you did it right. Once your article reaches 100 votes it is eligible for both publishing and rejection. 100 is an arbitrary number, it could be 5 votes or 500, whatever makes sense given your usercount. A supermajority of votes (67%) must vote to publish. An article which doesn't recieve 100 votes will "save" those votes it does recieve until the next week, when it will reappear in the RSS feed. An article with over 100 votes that does not have the supermajority for publishing is rejected, and goes back to the drawing board for review.

The collaborative process remains the same, but the collaborators must agree amongst themselves when to "version" an article and request it be published. So you and 10 other interested people could work on "Nonnegative_Integer" for a couple weeks, then discuss amongst yourselves that it's time to publish. An article can be edited many times a week, but can only be published once a week. Versions go up for review and voting on an "Edits" RSS feed, but publishing requires only a simple majority of yea votes, with no minimum vote limit. This means if you are the only person who votes for "Nonnegative_Integer" edits, they will go through. The publc sees only the published versions. This means vandalism becomes almost non-existent, and you can have a few super-users be able to unpublish vandalism etc.

You could have a lot of users who just enjoyed voting on articles. Maybe they are engineers who don't have the time to write an article themselves, but have a deep knowledge of math and enjoy reading and voting for one or two articles over coffee in the morning. The point of all this, however, is to ensure that anything which manages to get published has undergone significant peer review. Your idea about restricting priviliges to high-level users accomplishes the same thing; however it drastically reduces the pool of available editors, limiting collaboration and grinding the whole project to a halt. I believe my solution, while requiring more hours up front, will lead to a much more flexible and "wiki" like situation where interested parties can mini-peer review themselves. Of course this entire idea rests on the presumption that there are far more genuinely enthusiastic contributers than vandals and graffitti artists, so add that to your consideration.

Finally I'll point out that you can "jazz up" the presentation by exploring other types of wikis which could use this publishing model. If you can come up with enough good ideas about how to use more robust, accurate articles, you may be able to get the wikimedia foundation interested. Good luck with this project.


Multi-tiered Approach

[edit]

I have a genuine interest in working on this project. I would like to suggest that each article be composed of two sections: a "fact" section, which has passed some sort of peer review process and an "in debate" section, which chronicles new developments, contradictory evidence and competing theories. This way the project will be both a database of what we know as well as where we are going, without confusing the two. It would also make the peer review process more transparent, as all new posts would automatically go into the "in debate" section until confirmed.

As far as peer reviewing, I think we should have moderators that are experts in one particular field of science overseeing only the articles pertaining to their expertise. These experts would be the ones to decide what moves from "in debate" to "fact". Furthermore, I believe that membership should be required to post and should be contingent on some proof that the person will provide valuable information, such as a PhD in a scientific field or published papers. Again, members should only be allowed to post to articles in their field. As a neuroscientist, I have no business posting about mathematics. All this is pretty exclusionary, but if we want this to be a respected, reliable source of information I think it is necessary.

Finally, if we can get this up and running in an early version, we should submit the concept to scientific journals as letters to the editor or even brief communications to get the word out to the community and establish the project's legitimacy. I don't think it is unreasonable to think that we could get a grant from NIH or another government funding source to support this project and compensate coders and moderators for their time.

Right idea, wrong approach

[edit]

So sad that this has been sitting since 2006.

I want to talk about this like you would not believe. This is how we solve the climate crisis. been working on it for months.

If you look at the creation of wikipedia, it started out as nupedia, the professionally currated version of what turned into wikipedia. it did not work. I believe it took two weeks for wikipedia, the publicly edited version to outpace the invitation only version. Reality, it seems, would therefore act to poo poo barriers to entry for any such site.

As it stands, there still seem to be some major barriers to implementation of such a scheme. not insurmountable, but difficult. At least what i have found is that wikidata is geared for handling static data. science is not static. you would need to add wikiviariable to wikidata to make it work. Wikivariable probably needs a lot of computing power.

As i see it, the project really needs to pull from a lot of the wiki project community to work. wikiversity, WT:social. other aspects. huuuuuuuuuge project. chomping at the bit if you ever need someone to coordinate it in the proper spirit of all things wiki.